Considerable evidence has now accumulated that chronic hyperglycemia is associated with increased risk for the development of microvascular complications of insulin-dependent diabetes mellitus. Consequently, there has been an increased focus on the use of intensive insulin treatment strategies to improve glycemic control near those of non-diabetic individuals in order to decrease the incidence or slow the progression of complications. However, studies have shown that improved glycemic control leads to an increased frequency of severe hypoglycemia. Thus, there is a critical need to develop approaches to treatment of patients with insulin- dependent diabetes (IDDM) that lead to improved glycemic control while decreasing the risk for hypoglycemia. Our prior studies have shown that patients use their perceptions about their symptoms associated with hypoglycemia and hyperglycemia to make judgements about self-treatment However, findings from our current project highlight the inaccuracies of patient judgements about their blood glucose level based upon symptoms, and show that inaccuracy is in part due to loss of autonomic symptoms during hypoglycemia. Now, we propose to test a behavioral training program designed to increase patient ability to detect symptoms associated with sudden shifts in glucose level. The behavioral program will be tested under conditions of maximal importance, when intensive insulin treatment is being instituted to help patients achieve glycemic control near the range of individuals without IDDM. Eighty patients will be randomly assigned to one of two treatment conditions: 1) Blood Glucose Awareness Training (BGAT) plus intensive insulin treatment; 2) a control education program plus intensive insulin treatment. Patients will be evaluated using the hypoglycemic insulin clamp technique at baseline prior to randomization and four months after intensive insulin treatment and the education programs after been instituted. Our analyses will examine between-groups differences in accuracy of detection of hypoglycemia during the insulin clamp procedure. We will also compare the treatment groups in terms of frequency of blood glucose readings below 2.8 mmol/L, patient ability to detect hypoglycemia and hyperglycemia in actual field conditions, quality of life, and fear of hypoglycemia. If we find that the symptom awareness program leads to improved patient ability to detect acute glucose shifts, then we plan in the next phase of the project to follow patients over an extended period of time to examine whether BGAT alters the incidence of hypoglycemic events which are associated with loss of consciousness, seizures, or accidents, or events requiring assistance, thereby improving the safety of intensive insulin therapy.